KR20210056108A - Heat pump system for smart fish farms - Google Patents

Abstract

Disclosed is a heat pump system for fish farms to predict a failure thereof. According to the present invention, the heat pump system comprises: an aquaculture tank for aquaculture; a seawater tank receiving and supplying seawater to the aquaculture tank; a wastewater heat exchange unit installed in a drain line through which wastewater used in the aquaculture tank is discharged; a circulation unit circulating the wastewater; a first supply line receiving the seawater from the seawater tank and supplying the seawater, which is heated to a temperature higher than that of the seawater through the wastewater heat exchange unit, to the aquaculture tank; a hot water tank receiving the seawater from the seawater tank to store the seawater heated to a higher temperature than that of the seawater through the wastewater heat exchange unit and further increasing the temperature by a boiler; a second supply line receiving the seawater from the hot water tank and supplying the seawater to the aquaculture tank; a third supply line receiving the seawater from the seawater tank and supplying the seawater to the aquaculture tank; a recovery unit recovering the seawater from the first supply line to the seawater tank at a position in front of a hot water unit; a first sensor unit and a second sensor unit detecting the temperature and water storage amount of the aquaculture tank and the seawater tank, respectively; a third sensor unit, a fourth sensor unit, and a fifth sensor unit detecting the seawater temperature of the first supply line, the second supply line, and the third supply line; and a control unit performing calculation to supply the amount of seawater at the temperature required in the aquaculture tank with the lowest energy consumption. According to control of the control unit, the supply amounts of seawater through the first supply line, the second supply line, and the third supply line are adjusted.

Description

Heat pump system for smart farming water tank {HEAT PUMP SYSTEM FOR SMART FISH FARMS}

The present invention relates to a heat pump system for a smart aquaculture water tank, and more specifically, in a wastewater heat exchange unit, a rack and pinion gear (rack And pinion gear), a heat pump system for a smart aquaculture water tank adopting a structure that can be attached to a device using a structure, a first sensor unit and a second sensor unit for detecting the temperature and storage amount of the aquaculture tank and seawater tank, respectively, and a first supply line, Using the second supply line and the third sensor unit, the fourth sensor unit, and the fifth sensor unit for detecting the seawater temperature of the third supply line, the control unit determines the amount of seawater at the temperature required by the fish tank at the lowest energy. A heat pump system for a smart aquaculture tank capable of calculating a method for supplying by consumption and controlling the amount of seawater supplied through the first supply line, the second supply line, and the third supply line according to the control of the control unit. Through the monitoring unit, based on the image information acquired from the photographing camera and the sensing information sensed by the sensor unit, the failure of the heat pump system for aquaculture tank is predicted in advance through comparison with the image information and sensing information at the time of failure. It is possible to do, it relates to a heat pump system for a smart aquaculture tank.

Aquaculture is the breeding of fish in aquaculture, and there are methods such as pond farming, flowing water farming, cage farming, circulation filtering farming, and stocking and repo farming. Such aquaculture industry has remained at the level of the traditional industry until recently, but in recent years, it is transforming into a high-tech industry in which the fourth industrial revolution technologies such as ICT, robots, artificial intelligence, and big data are fused. Tech Cast, an American corporate consulting firm, predicts that the market for aquaculture technology will reach 223 trillion won by 2025.

Meanwhile, due to rising oil prices and manpower shortages, the offshore fishing industry has entered the red ocean, and as described above, the aquaculture industry is growing rapidly, but when farming on land, oil boilers have been used a lot to raise the water temperature in winter. However, due to the volatility of oil prices, the fish family had a burden on the use of oil boilers.

In this situation, the heat pumps of farms that are recently being supplied are attracting attention as alternative energy facilities, and the government also provides subsidiary projects for facility costs when the boiler facilities are removed and heat pump facilities are installed to reduce the supply of duty-free oil and improve environmental problems. Had been implemented. Through this, the demand for heat pumps in fisheries farms has rapidly increased.

As a related art of the heat pump of a farm related to the related art, in Korean Patent Publication No. 10-1401453 (name of the invention: HEAT PUMP APPARATUS OF FISH FARMS, registration date: May 23, 2014), "The present invention is Regarding the heat pump device of a farm, it is sealed with a casing with a sound insulation member attached to a plurality of pumps, and a separate shielding member is used to seal the exposed hole of the pipe, and the pump casing is positioned above a certain height to prevent seawater and rainwater. It is an object of the present invention to provide a heat pump device for a farm that reduces the possibility of corrosion and suppresses noise and vibration by forming a pump stand in a vibration-proof structure. It has the effect of removing the factors that inhibit the growth of fish and shellfish in aquaculture and improving the noise environment in the farm. Second, it prevents the inflow of seawater, base and rainwater by configuring a casing in the seawater pump installed together with the heat pump in the farm. Started "Technology that has the advantage of preventing seawater from splashing or flowing into the pump casing when fish and shellfish are withdrawn from the aquaculture tank by configuring the pump casing on the upper surface of a separate support, which extends the life of the pump and reduces failures." Are doing.

On the other hand, in the Republic of Korea Patent Publication No. 10-1632202 (name of the invention: floor coil heating method heat pump system for aquaculture water tank, registration date: 2016.06.15), "A floor coil heating method aquaculture water tank heat pump system is disclosed. The disclosed heat pump system for a bottom coil heating method aquaculture tank is a tank bottom heating coil disposed at the bottom of the culture tank to heat water in the culture tank; a heated circulation fluid capable of transferring heat while passing through the tank bottom heating coil A heat storage tank in which the heated circulating fluid is circulated and heated while being circulated between the heat storage tank and the heat storage tank; a waste heat tank containing waste heat and containing waste water discharged from the aquaculture tank; and the waste heat heat pump. Including a heated heat exchanger capable of heating the waste circulating fluid circulating between the waste heat while exchanging heat with the wastewater discharged from the waste heat tank; and an electrolysis tank for backwashing capable of backwashing the heated heat exchanger; and the During the heating operation for transferring heat to the aquaculture tank, the waste water contained in the waste heat tank passes through the heated heat exchanger and exchanges heat with the waste circulation fluid circulating between the waste heat heat pump and the heated heat exchanger, so that the waste circulation fluid Is heated, and the waste circulation fluid heated while passing through the heated heat exchanger passes through the waste heat heat pump and exchanges heat with the heated circulation fluid accommodated in the heat storage tank through the waste heat heat pump, thereby heating the heated circulation fluid, During a backwash operation in which the heated circulating fluid accommodated in the heat storage tank after being heated via the waste heat heat pump heats the inside of the aquaculture tank while passing through the tank bottom heating coil and backwash the heated heat exchanger, the electrolysis for backwashing Disclosed is a technique, characterized in that the heated heat exchanger is backwashed while the external washing water electrolyzed while passing through a bath passes through the heated heat exchanger.

However, in the above-described two patent documents, the first sensor unit and the second sensor unit respectively detecting the temperature and the storage amount of the aquaculture tank and the seawater tank, and the seawater temperature of the first supply line, the second supply line, and the third supply line. Using the third sensor unit, the fourth sensor unit, and the fifth sensor unit that detects, the control unit calculates a method for supplying the seawater amount of the temperature required by the aquarium water tank at the lowest energy consumption, and is used for the control of the control unit. Therefore, through the heat pump system for a smart aquaculture tank, and further, through a monitoring unit, it is possible to control the amount of seawater supplied through the first supply line, the second supply line, and the third supply line, and image information and sensors obtained from the photographing camera Based on the sensing information sensed by the department, it is possible to predict the failure of the heat pump system for aquaculture tank in advance through comparison with the image information and sensing information at the time of failure. There was no initiation.

1. Korean Patent Publication No. 10-1401453 (Name of invention: HEAT PUMP APPARATUS OF FISH FARMS, registration date: May 23, 2014) 2. Republic of Korea Patent Publication No. 10-1632202 (Name of invention: floor coil heating method aquaculture water tank heat pump system, registration date: 2016.06.15)

The present invention was created to solve the above-described problem, the present invention is to a mechanism using a rack and pinion gear so that it is possible to clean a plurality of tubes while rotating a plurality of nozzles without cleaning one by one in a wastewater heat exchange unit. Its purpose is to provide a heat pump system for a smart aquaculture water tank that adopts a structure that can be attached.

In addition, another object of the present invention is to measure the seawater temperature of the first and second sensor units, and the first supply line, the second supply line, and the third supply line, respectively, for detecting the temperature and storage amount of the aquaculture tank and the seawater tank. Using the third sensor unit, the fourth sensor unit, and the fifth sensor unit to detect, the control unit calculates a method for supplying the seawater at the temperature required by the fish farm water tank at the lowest energy consumption, and according to the control of the control unit. , It is possible to control the supply amount of seawater through the first supply line, the second supply line, and the third supply line, to provide a heat pump system for a smart aquaculture tank.

In addition, the present invention, through a monitoring unit, based on the image information acquired from the photographing camera and the sensing information sensed by the sensor unit, through comparison with the image information and sensing information at the time of failure in the past, the heat pump system for aquaculture tank Another object is to provide a heat pump system for a smart aquaculture water tank, capable of predicting the failure in advance.

In addition, in the monitoring unit of the bottom pump system for a smart aquaculture tank, in particular, a heat pump system for a smart aquaculture tank with an artificial intelligence camera function capable of minimizing the blurry area and converting a sharp high-resolution image from a low-resolution image of a shooting camera. Another purpose is to provide.

A heat pump system for a smart aquaculture water tank according to the present invention for achieving the above object comprises: a farm water tank for aquaculture; A seawater tank for receiving seawater and supplying seawater used for the aquaculture tank; A wastewater heat exchange unit installed in a drain line through which the wastewater used in the farm water tank is discharged; A circulation unit for circulating the wastewater; A first supply line for receiving seawater from the seawater tank and supplying the seawater raised to a temperature higher than the seawater through the wastewater heat exchange unit to the aquaculture tank; A hot water tank capable of receiving seawater from the seawater tank, storing the seawater raised to a temperature higher than the seawater through the wastewater heat exchange unit, and further increasing the temperature through a boiler; A second supply line for receiving seawater from the hot water tank and supplying it to the aquaculture tank; A third supply line for receiving seawater from the seawater tank and supplying the seawater to the aquaculture tank; And a recovery unit for recovering seawater from a first supply line to the seawater tank at a position in front of the hot water unit, wherein the wastewater heat exchange unit includes a heat pump comprising a shell and a plurality of tubes located in the shell, and the It adopts a structure that can be attached to a mechanism using a rack-and-pinion gear to enable cleaning while rotating a plurality of nozzles without cleaning a plurality of tubes one by one.

In addition, a first sensor unit and a second sensor unit respectively detecting the temperature and storage amount of the aquaculture tank and the seawater tank, and a third sensor detecting the seawater temperature of the first supply line, the second supply line, and the third supply line. A unit, a fourth sensor unit, and a fifth sensor unit; And a control unit that calculates the amount of seawater at a temperature required by the aquaculture tank with minimum energy consumption, and further comprises a first supply line, a second supply line, and a third supply line according to the control of the control unit. The amount of seawater supplied through the supply line can be adjusted.

In addition, the farm water tank, the seawater tank, the waste water heat exchange part, the drain line, the circulation part, the first supply line, the boiler, the hot water tank, the second supply line, the third supply line, and the Each of the collection units is monitored, but a monitoring unit including a photographing camera and a sensor unit for monitoring a still image at a predetermined period for power saving, wherein the monitoring unit further includes, in the monitoring unit, image information obtained from the photographing camera and the sensor Based on the sensing information sensed by the unit, it is possible to predict the failure of the heat pump system for the aquaculture water tank in advance by comparing the image information and the sensing information at the time of failure beforehand.

In addition, it is installed between the aquaculture tank and the wastewater heat exchange unit, to prevent failure of the wastewater heat exchange unit and to increase heat exchange efficiency, a foreign material removal unit for removing foreign substances from the wastewater discharged from the fish farm water tank; further comprises I can.

In addition, the circulation unit is driven by a motor controlled by a motor control unit to drive the wastewater pump in connection with a signal for an inlet pressure corresponding to the excessive discharge of the wastewater by a pressure sensor installed in the wastewater heat exchange unit. desirable.

In addition, a data processing unit that increases the resolution of the image information on the still image acquired by the photographing camera; further comprising, the data processing unit, by using the stored artificial intelligence variables, acquired with respect to the object monitored by the monitoring unit. The process of increasing the resolution of at least two or more image data from at least two or more cameras is performed, but the artificial intelligence variables are artificial intelligence learning of the discriminator network and the generator network for at least two or more image data. Learn with an artificial intelligence network that repeats the process of intelligent learning.

According to the heat pump system for a smart aquaculture water tank according to the present invention,

First, the wastewater heat exchange unit includes a heat pump consisting of a shell and a plurality of tubes located in the shell, and a mechanism using a rack and pinion gear to enable cleaning while rotating a plurality of nozzles without cleaning a plurality of tubes one by one. Adopt a structure that can be attached to.

Second, a first sensor unit and a second sensor unit respectively detecting the temperature and storage amount of the aquaculture tank and the seawater tank, and a third sensor unit detecting the seawater temperature of the first supply line, the second supply line, and the third supply line, Using the fourth sensor unit and the fifth sensor unit, the control unit calculates a method for supplying the seawater amount of the temperature required by the fish farm tank at the lowest energy consumption, and according to the control of the control unit, the first supply line and the first supply line It is possible to control the amount of seawater supplied through the 2 supply line and the 3rd supply line.

Third, through the monitoring unit, based on the image information acquired from the photographing camera and the sensing information sensed by the sensor unit, the failure of the heat pump system for the aquaculture tank is detected through comparison with the image information and sensing information at the time of failure. It is possible to predict in advance.

Fourth, in the monitoring unit of the bottom pump system for a smart aquaculture tank, the heat pump system for a smart aquaculture tank has an artificial intelligence camera function capable of minimizing the blurry area and converting a sharp high-resolution image from a low-resolution image of a photographing camera, in particular. Can provide.

는 샤프하면서도 고해상도를 갖는 이미지가 만들어질 수 있다. 따라서 상대적으로 저렴한 가격, 또는 중고의 저해상도 카메라 렌즈를 통하여 촬영하더라도, 고해상도 이미지를 획득할 수 있어서, 양식 수조용 히트펌프 시스템 고장의 사전 예측의 정확도를 높이는 것이 가능하다. That is, a generated image converted through the creator network 230 to a low-resolution combined image made of two or more low-resolution images.

Can create sharp and high-resolution images. Therefore, it is possible to obtain a high-resolution image even when photographing through a relatively inexpensive or used low-resolution camera lens, and thus, it is possible to increase the accuracy of pre-prediction of failure of the heat pump system for aquaculture tank.

1 is a schematic block diagram of a heat pump system for a smart aquaculture water tank according to an embodiment of the present invention.
2 is a view for explaining the automatic cleaning of the heat pump in the heat pump system for a smart aquaculture water tank according to an embodiment of the present invention.
3 is a block diagram of an artificial intelligence camera system 100 according to an embodiment of the present invention.
4 is a diagram illustrating a method of learning artificial intelligence with an artificial intelligence network in an artificial intelligence camera system according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Prior to this, terms or words used in the specification and claims should not be construed as being limited to their usual or dictionary meanings, and the inventors appropriately explain the concept of terms in order to explain their own invention in the best way. Based on the principle that it can be defined, it should be interpreted as a meaning and concept consistent with the technical idea of the present invention.

Accordingly, the embodiments described in the present specification and the configurations shown in the drawings are only the most preferred embodiment of the present invention, and do not represent all the technical spirit of the present invention. It should be understood that there may be equivalents and variations.

(Smart Aquaculture Water Tank Heat Pump System)

1 is a schematic block diagram of a heat pump system for a smart aquaculture water tank according to an embodiment of the present invention.

As shown in FIG. 1, the heat pump system for a smart aquaculture water tank according to an embodiment of the present invention includes a farm water tank 10, a sea water tank 20, a wastewater heat exchange part 30, a circulation part 90, Hot water tank 50, first supply line 40, second supply line 42, third supply line 44, recovery unit 60, control unit (not shown), and monitoring unit (not shown), etc. Includes.

The fish farm tank 10 refers to a tank for aquaculture, and the shape of the tank can be a natural tank as well as an artificial tank. The seawater tank 20 is a tank that receives seawater and supplies seawater to be used in the aquarium tank.

The wastewater heat exchange unit 30 is a structure installed in a drainage line through which wastewater used in the fish farm water tank 10 is discharged. The circulation unit 90 is configured to circulate wastewater, and a motor to drive the wastewater pump in connection with a signal for inlet pressure corresponding to excessive discharge of wastewater by a pressure sensor installed in the wastewater heat exchange unit 30 described above. It is driven by a motor controlled by the control unit.

The first supply line 40 receives seawater from the seawater tank 20, passes through the wastewater heat exchange unit 30, and supplies the seawater that is heated to a higher temperature than the seawater to the aquaculture tank 10. The hot water tank 50 receives seawater from the seawater tank 20 and stores the seawater that has been heated higher than the seawater through the wastewater heat exchange unit 30, and supplies the additionally heated seawater through a separate boiler 55. It is possible to supply to the aquaculture tank 10 through the line 42. The third supply line 44 may receive seawater from the seawater tank 20 and supply seawater to the aquaculture tank 10.

In addition, the recovery unit 60 is configured to recover seawater from the first supply line 40 to the seawater tank 20 at a position in front of the hot water tank 50.

In addition, the first sensor unit and the second sensor unit respectively detecting the temperature and the storage amount of the aquaculture tank 10 and the seawater tank 20, and the seawater temperature of the first supply line, the second supply line, and the third supply line. It further includes a third sensor unit, a fourth sensor unit, and a fifth sensor unit.

In addition, the control unit (not shown) is a configuration that calculates to supply the amount of seawater at the temperature required by the aquaculture tank 10 with the lowest energy consumption. According to the control of the control unit, the amount of seawater supplied through the valves present in the first supply line 40, the second supply line 42, and the third supply line 44, for example, an electronic control valve, can be appropriately adjusted. I can. This minimizes energy consumption and maximizes energy efficiency. Here, by using a flow sensor, it is also possible to estimate the temperature by reducing the number of temperature sensors by using the calculation formula that the mass flow rate of the heat source * the temperature difference value at the inlet and outlet of the heat source = the mass flow rate of the working fluid * the temperature difference value at the inlet and outlet of the working fluid. It is possible.

In addition, it may further include a monitoring unit (not shown). The monitoring unit, the farm water tank 10, the seawater tank 20, the waste water heat exchange unit 30, the drain line 46, the circulation unit 90, the first supply line 40, the boiler 55, the hot water tank 50, the second supply line 42, the third supply line 44, and the recovery unit 60 are monitored, respectively, but to save power, they are monitored as still images at predetermined intervals. Of course, real-time monitoring is possible with video, but still images are taken at predetermined intervals or at predetermined times for power saving.

Further, the monitoring unit includes a photographing camera and a sensor unit, and in the monitoring unit, based on the image information acquired from the photographing camera and the sensing information sensed by the sensor unit, the , Predict failure of the heat pump system for aquaculture tank. Previously, image information and sensing information at the time of failure are stored in a database (not shown), and a set of feature points or preceding patterns in image information or sensing information occurring before an accident are stored through image processing in advance. It is possible.

In addition, a foreign material removal unit 70 may be further included. A configuration that is installed between the farm tank 10 and the wastewater heat exchange part 30 to remove foreign substances from the wastewater discharged from the farm tank 10 so as to prevent failure of the wastewater heat exchange part 30 and increase heat exchange efficiency. As such, it may be in the form of a filter. Foreign substances may be seaweeds that can be found in seawater or carcasses of dead fish among fish in aquaculture farms.

In addition, a data processing unit (not shown) is further provided. The data processing unit is configured to increase the resolution of image information on a still image acquired from a photographing camera, and acquires an object monitored by the above-described monitoring unit, for example, a part of the wastewater heat exchange unit, using pre-stored artificial intelligence variables. A process of increasing the resolution of at least two or more image data from at least two or more shooting cameras is performed, but the artificial intelligence variables are artificial intelligence learning of the discriminator network and the generator network for at least two or more image data. Learn with an artificial intelligence network that repeats the process of intelligent learning.

(Automatic cleaning of heat pump)

The biggest cause of failure in the heat pump system for smart aquaculture water tank is the shutdown of the heat pump due to sludge. In order to prevent such downtime, failure prediction data by IoT sensing is accumulated.

The shell and tube of the shell & tube heat pump are equipped with IoT sensors for monitoring to accumulate operation data. The measured value from the monitoring sensor of each part is saved as big data and sent to the monitoring program to check whether an event has occurred. When an abnormal signal is detected, it is immediately analyzed. It is also possible to warn through an alarm that it is necessary to actively clean the tube if the measured value higher than the preset set value is continued.

If a tube cleaning warning occurs, immediately stop the pump and perform cleaning. Rather than cleaning multiple tubes one by one, attach a device using rack and pinion gears (rack and pinion gear), and attach a device using rack and pinion gears to complete work faster while rotating multiple nozzles. To have a structure.

As shown in Figure 2, sludge sensing is performed as follows. It accumulates long-term data measured by the sensor at each measurement point. The sensor temperature in a specific location rises, and the event mode is switched by detecting the temperature rise. A warning is sent by determining the similarity of the analyzed pattern. Furthermore, it adopts a structure capable of attaching a rotary tube sludge cleaner.

In short, the wastewater heat exchange unit includes a heat pump consisting of a shell and a plurality of tubes located in the shell, and a rack and pinion gear (rack and pinion gear (rack and It is characterized by adopting a structure that can be attached to a device using (pinion gear).

(Artificial intelligence camera system configuration)

3 is a block diagram of an artificial intelligence camera system 100 according to a first embodiment of the present invention. As shown in FIG. 3, a data photographing unit 110, a data storage unit 120, and a data processing unit 130 are included.

Here, the data photographing unit 110 is a configuration for acquiring image data for a subject photographed by a heat pump system for a smart water tank, and is usually implemented by a camera or the like. Here, the data photographing unit 110 includes a camera module 101 to obtain image data for a subject, and may acquire two or more image data for a subject when photographing with one camera module 101. . Here, two or more image data may be photographed at slightly different locations. Alternatively, two or more image data may be acquired by photographing using two or more camera modules.

In addition, the data storage unit 120 is a component for storing image data and artificial intelligence variables acquired by the data photographing unit 100. Here, the image data or artificial intelligence variables stored in the data storage unit 120 may be stored in a volatile memory that is received and stored in real time and then deleted.

In addition, the data processing unit 130 is a component that increases the resolution of the stored image data using stored artificial intelligence variables. Here, the artificial intelligence variables stored in the data storage unit 120 and used to increase the resolution of the image in the data processing unit 130 are the process of artificial intelligence learning the discriminator network for at least two or more image data, and As an artificial intelligence network (Generative Adversarial Network (GAN)) that learns two networks of the generator network 230 and the discriminator network 260 hostilely by repeating the process of artificial intelligence learning the generator network. Artificial intelligence learning is possible.

(Generate combined image)

4 is a diagram illustrating a method of learning artificial intelligence with an artificial intelligence network in an artificial intelligence camera system according to an embodiment of the present invention.

As shown in FIG. 4, in order to implement sharp edges when generating a high-resolution image from a low-resolution image, the artificial intelligence network of the present invention connects two networks of the generator network 230 and the discriminator network 260 to each other. It can be learned hostile.

In addition, the low-resolution images z1, ..., zn (210) obtained by photographing by the data photographing unit 110 are acquired by the following method to obtain the combined image z (220). When the size of the low-resolution images z1 and z2 is W (horizontal) x H (vertical) x C (number of channels), the combined image z of these two images can be defined as an image having a dimension of W × H × 2C.

That is, it becomes as shown in Equation 1 below.

Where i=1,2, ..., W, j=1,2, ...., H, k=1,2,...,2C, W, H, and C are natural numbers to be.

Alternatively, you can create a combined image z like this:

Where i=1,2, ..., W, j=1,2, ...., H, k=1,2,...,C, W, H, and C are natural numbers to be.

In addition, a method of combining two low-resolution images can be implemented by various modifications within the context of the present invention. For the above coupling method, n=2 was set for convenience, but the present invention is not limited thereto. It goes without saying that it can be expanded to n=3,4,... or the like as much as possible within the content of the present invention.

For example, in the case of n=r×r (e.g., n=4=2×2), the above combining method increases the size of the image using a pixel shuffler in addition to increasing the number of channels. May be. The pixel shuffler expands an image of size W × H × C into an image of size rW × rH × C.If r=2, for example, the pixel at position (i, j) in the z1 image is r × r (e.g. :2×2), but the corresponding pixels in z2, z3, and z4 are combined as shown in Equation 4 below. That is, in the case of four images z1, z2, z3, and z4, a low-resolution combined image z is generated by Equation 3.

Where i=1,2, ..., W, j=1,2, ...., H, k=1,2,...,C, W, H, and C are natural numbers to be.

(Artificial Intelligence Learning)

As described above, after acquiring the low-resolution combined image z, artificial intelligence training is performed in the following manner.

라고 하고, N개의 저해상도 결합 이미지를 나타내는 데이터 z를 포함하는 저해상도 결합 이미지들의 집합을

라고 할 때, 생성자 네트워크의 인공지능 변수 벡터

는 고정시킨 상황에서, 저해상도 결합 이미지를 나타내는 데이터 z에 대하여, 생성자 네트워크(230)를 거쳐 생성이미지

(240)를 얻고 이를 다시 판별자 네트워크(260)로 거치며 얻은 최종 변환된 값

을 0(거짓)(270)으로 되게 하고, 동시에 고해상도 이미지 x들을 판별자 네트워크(260)로 입력하였을 때의 변환값

는 1(참)(270)이 되도록, 즉 수학식 5를 포함하는 비용함수(cost function)을 최소화 되도록, 판별자 네트워크의 인공지능 변수 벡터(θ)를 학습시킨다. A set of high-resolution images including N high-resolution images x

And a set of low-resolution combined images including data z representing N low-resolution combined images

Is, the vector of artificial intelligence variables of the generator network

Is a generated image through the generator network 230 with respect to the data z representing the low-resolution combined image in a fixed situation.

The final converted value obtained by obtaining (240) and passing it back to the discriminator network (260)

Is set to 0 (false) 270, and at the same time, the converted value when high-resolution images x are input to the discriminator network 260

Is 1 (true) 270, that is, to minimize the cost function including Equation 5, the artificial intelligence variable vector θ of the discriminant network is trained.

는 고정시킨다. 이것은 판별자가 저해상도 결합 이미지로부터 생성된 이미지는 판별자 네트워크(260)를 거치면서 거짓으로 판명나게 하고 원래의 고해상도 이지미는 판별자 네트워크(260)를 거치면서 참으로 판명나게 학습시키는 과정이다. In other words, the artificial intelligence variable vector of the generator network 230

Is fixed. This is a process in which an image generated from a low-resolution combined image by a discriminator turns out to be false while passing through the discriminator network 260, and the original high-resolution image is learned to turn out to be true while passing through the discriminator network 260.

(240)를 판별자 네트워크(260)를 거치면서 참이 되게 하는 인공지능 학습 과정이다. 즉 생성자 네트워크(230)는 판별자를 속이는 역할을 하게 하는 인공지능 학습 과정이다. Next, so that the cost function including Equation (5) is minimized, that is, the artificial intelligence variable vector (θ) of the discriminator network 260 is fixed, and generated through the generator network 230 of N low-resolution combined images z image

It is an artificial intelligence learning process that makes (240) true while passing through the discriminator network (260). That is, the generator network 230 is an artificial intelligence learning process that plays the role of deceiving the discriminator.

를 최적화되게 학습시킬 수 있다. 이로부터 변환된 생성이미지

는 고해상도 이미지를 샤프하게 복원할 수 있다. In this way, the artificial intelligence variable vector through the generator network 230 from the low-resolution data z through the artificial intelligence learning network

Can be trained to be optimized. Generated image converted from this

Can sharply restore high-resolution images.

는 상기 저해상도 결합 이미지를 나타내는 데이터 z들의 생성자 네트워크를 거친 생성이미지

를 다시 판별자 네트워크로 거쳐 얻은 최종 변환된 이미지이다. here,

Is a generated image through a generator network of data z representing the low-resolution combined image

Is the final transformed image obtained by passing through the discriminator network again.

이 해당 고해상도 이미지의 픽셀

와 같아지도록 제약하는 수학식 6으로 표현되는 비용함수를 같이 추가로 포함시켜 학습시킬 수도 있다. Further, during the artificial intelligence learning process, each pixel of the image learned through the generator network 260 is to generate the low-resolution image z in the cost function including Equation 6

The pixels of this corresponding high-resolution image

It is also possible to additionally include and learn the cost function represented by Equation 6 that restricts to be equal to.

Here, W and H denote the width and height of the image, respectively.

Furthermore, it is possible to learn by adding a constraint expressed by Equation 7 instead of Equation 6.

는 고해상도 이미지 x에 대한 VGG(Visual Geometry Group)의 네트워크의 레이어 l에서 i번째 특징지도(feature map)의 j번째 원소를 나타낸다. 한편,

는 저해상도 결합 이미지 z를 생성자 네트워크(230)를 통하여 변환되어 생성된

의 VGG의 네트워크의 레이어 l에서 i번째 특징지도의 j번째 원소를 나타낸다. 즉, 수학식 7은 저해상도 결합 이미지로부터 변환된 이미지의 콘텐츠가 고해상도의 콘텐츠를 갖게 되도록 하는 제약조건일 수 있다. here,

Denotes the j-th element of the i-th feature map in the layer l of the VGG (Visual Geometry Group) network for the high-resolution image x. Meanwhile,

Is generated by transforming the low-resolution combined image z through the generator network 230

Represents the j-th element of the i-th feature map in the layer l of the VGG network of. That is, Equation 7 may be a constraint that allows content of an image converted from a low-resolution combined image to have high-resolution content.

For reference, VGG is a convolutional neural network (CNN) model proposed by Oxford University and is a very successful model for image classification. The VGG network of the present invention may select the VGG19 model. The VGG19 model consists of 16 convolutions, 5 max pooling, and 3 fully connected (FC) models.

는 고해상도 이미지 x 들이 분포되어 있는 다양체(submanifold)에 속하게 할 수 있다. 즉 샤프하면서도 고해상도를 최대로 복원하는 이미지를 생성할 수 있다. In this way, the generator network 230 is learned from the low-resolution combined image created by acquiring two or more low-resolution images and combining them through the learning process of the artificial intelligence network, and the generated image converted therefrom.

Can belong to a submanifold in which high-resolution images x are distributed. In other words, it is possible to create an image that is sharp and restores high resolution to the maximum.

As described above, although the present invention has been described by the limited embodiments and drawings, the present invention is not limited thereto, and the technical spirit of the present invention and the following by those of ordinary skill in the technical field to which the present invention pertains. It goes without saying that various modifications and variations are possible within the equal range of the claims to be described.

(Explanation of code)

10...aquaculture tank

20...sea tank

30...wastewater heat exchange section

40...first supply line

42...second supply line

44...3rd supply line

46...drain line

50...hot water bath

55...boiler

60...recovery

70... foreign matter removal section

80...control valve

90...circulation

100...Artificial intelligence camera system

101...Camera module

110...data capture department

120...data storage

130...data processing unit

Claims (6)

Aquaculture tank for aquaculture;
A seawater tank for receiving seawater and supplying seawater used for the aquaculture tank;
A wastewater heat exchange unit installed in a drain line through which the wastewater used in the farm water tank is discharged;
A circulation unit for circulating the wastewater;
A first supply line for receiving seawater from the seawater tank and supplying the seawater raised to a temperature higher than the seawater through the wastewater heat exchange unit to the aquaculture tank;
A hot water tank capable of receiving seawater from the seawater tank, storing the seawater higher than the seawater through the wastewater heat exchange unit, and further increasing the temperature through a boiler;
A second supply line for receiving seawater from the hot water tank and supplying it to the aquaculture tank;
A third supply line for receiving seawater from the seawater tank and supplying the seawater to the aquaculture tank; And
Includes; a recovery unit for recovering seawater from the first supply line to the seawater tank at a position in front of the hot water unit,
The wastewater heat exchange unit includes a heat pump consisting of a shell and a plurality of tubes located in the shell, and a rack and pinion gear is used to clean the plurality of tubes while rotating a plurality of nozzles without cleaning one by one. Characterized in that adopting a structure that can be attached to,
Heat pump system for aquaculture tank.
The method of claim 1,
A first sensor unit and a second sensor unit respectively detecting the temperature and storage amount of the aquaculture tank and the seawater tank, and a third sensor unit detecting the seawater temperature of the first supply line, the second supply line, and the third supply line, A fourth sensor unit and a fifth sensor unit; And
A control unit that calculates to supply the amount of seawater at a temperature required by the aquaculture tank with the lowest energy consumption; further comprising,
According to the control of the control unit, characterized in that the amount of seawater supplied through the first supply line, the second supply line, and the third supply line is adjusted,
Heat pump system for aquaculture tank.
The method of claim 2,
The farm water tank, the seawater tank, the waste water heat exchange part, the drain line, the circulation part, the first supply line, the boiler, the hot water tank, the second supply line, the third supply line, and the recovery part Each monitoring unit, including a photographing camera and a sensor unit, which monitors as a still image at a predetermined period for power saving, and further includes,
In the monitoring unit, based on the image information acquired by the photographing camera and the sensing information sensed by the sensor unit, the heat pump system for the aquaculture tank is compared with the image information and sensing information at the time of failure. Characterized in that the failure is predicted in advance,
Heat pump system for aquaculture tank.
The method of claim 3,
A foreign matter removal unit installed between the aquaculture tank and the wastewater heat exchange unit to remove foreign substances from the wastewater discharged from the aquaculture tank so as to prevent failure of the wastewater heat exchange unit and increase heat exchange efficiency.
Heat pump system for aquaculture tank.
The method of claim 4,
The circulation unit is driven by a motor controlled by a motor control unit to drive the wastewater pump in connection with a signal for an inlet pressure corresponding to the excessive discharge of the wastewater by a pressure sensor installed in the wastewater heat exchange unit. doing,
Heat pump system for aquaculture tank.
The method of claim 5,
A data processing unit that increases the resolution of image information for the still image acquired by the photographing camera; further comprising,
The data processing unit performs processing to increase the resolution of at least two or more image data from at least two or more photographing cameras acquired with respect to an object monitored by the monitoring unit by using the stored artificial intelligence variables, wherein the artificial intelligence variable They are characterized in that they learn with an artificial intelligence network that repeats the process of artificial intelligence learning of the discriminator network and artificial intelligence learning of the generator network for at least two or more image data,
Heat pump system for aquaculture tank.

Images